In addition to risk factors measured on an individual level, such as diet, family history of cancer and cancer screening behaviors, factors measured on a group or area level may influence health. To investigate associations between colorectal cancer incidence in women, socioeconomic status and environmental risk factors, we propose to link publicly-available data sets with risk factor information gathered as part of a recently completed case-control study. In particular, we hypothesize that (1) area variations in colorectal cancer incidence are explained by individual-level socioeconomic status indicators and other established risk factors including screening history; and (2) contaminants in drinking water - especially trihalomethanes, nitrates and pesticides - are associated with increased colorectal cancer risk. To address these hypotheses, we propose to geocode (assign a latitude and longitude to) 833 female colorectal cases diagnosed in 1997-2000 and 2,123 population-based controls. Comprehensive risk factor questionnaire information from structured telephone interviews is available on all cases and controls. By geocoding these study participants, we can place each case and control on a map. We can then overlay this map with information concerning socioeconomic status (Census) and water quality (Wisconsin Departments of Natural Resources and Agriculture, Trade, and Consumer Protection). Multi-level modeling and spatial analysis will be employed to take full advantage of the opportunities of this rich combined data source. This project will also provide the foundation for numerous future linkages between individual-level risk factor histories and publicly available datasets as well as the application of statistical techniques for assigning exposure to geographically defined, potentially carcinogenic compounds.